Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
Hodgson E. Modern toxicology [2004].pdf
6.63 Mб





Range of



Observable Range

Figure 24.2 Dose-responsecurve, with emphasis on the shape of thedose-responsefunction below the experimentally observable range and therefore the range of inference where people are realistically exposed.

U-shapedeffects can be explained in terms of homeostatic adjustments or overcorrections in the operation of feedback mechanisms. Examples of studies with data fitting aU-shapedcurve include the hormetic effect of organic lead on body growth in rats (Cragg and Rees, 1984) and peripheral nerve conduction velocity in children at low doses (Ewert et al., 1986). Similar relationships have been observed with alcohol and nicotine in humans. It has been proposed that because thresholds are inherent inU-shapeddose-responsecurves, the linearno-thresholdextrapolation method is not an appropriate approach for regulating hormetic agents. The current risk assessment paradigm used by US EPA and other federal agencies does not conflict with the concept of hormesis, but it has been proposed that the risk assessor’s analyzes make an active consideration of the data and the application of that data in the low dose portion of thedose-responsecurve for hormetic agents.


The noncancer risk assessment process assumes a threshold. For many noncarcinogenic effects, protective mechanisms are believed to exist that must be overcome before an adverse effect is manifested. At the cellular level for some toxicant, a range of exposures exists from zero to some finite value that can be tolerated by the organism with essentially no chance of expression of adverse effects. The aim here in risk assessment is to identify the upper bound of this tolerance range (i.e., the maximum subthreshold level). This approach involves obtaining the no observed adverse effect level. NOAEL is the highest dose level thatdoes not produce a significant elevated increase in an adverse response. Significance refers to biological and statistical criteria and is dependent on dose levels tested, number of animals, background incidence in the unexposed control groups. Sometimes there is insufficient data to arrive at a NOAEL, and a LOAEL (lowest observed adverse effect level) is derived. The NOAEL is the key datum obtained from the study of thedose-responserelationship. The NOAEL is used to calculate reference doses (RfD) for chronic oral exposures and reference concentrations (RfC) for chronic inhalation exposures as per EPA. Other agencies, such as the ATSDR and WHO, use the NOAEL to calculateminimum risk levels (MRLs) andacceptable daily intakes (ADI). The US EPA describes the RfD as an estimate, with uncertainty spanning an order of magnitude, of a daily exposure to the human population, including


sensitive subgroups, that is likely to be without appreciable deleterious effects during a lifetime. In deriving reference doses, ADIs, or MRLs, the NOAEL is divided by uncertainty factors (UF) as per EPA (EPA, 1989) and ATSDR (ATSDR, 1993) and by modifying factors (MF) as per EPA:

RfD =



























The calculated RfD or RfC is based on the selected critical study and selected critical end point. The risk assessor may obtain numerous studies where the toxicant may have more than one toxic end point, and thus there may be many NOAELs to choose from the literature. In some instances poor data quality may be used to exclude those end points from consideration. Also at issue is the determining what is considered an adverse effect, and this has been summarized with a few examples in Table 24.2. In sum, the MRL or RfD is based on the less serious effects and no serious effects. The following are example effects not used in obtaining a NOAEL: decrease in body weight less than 10%, enzyme induction with no pathologic changes, changes in organ weight with no pathologic changes, increased mortality over controls that is not significant (p > 0.05), and hyperplasia or hypertrophy with or with out changes in organ weights.

24.3.1Default Uncertainty and Modifying Factors

Most extrapolations from animal experimental data in the risk assessments require the utilization of uncertainty factors. This is because we are not certain how to extrapolate across species, with species for the most sensitive population, and across duration. To account for variations in the general population and to protect sensitive subpopulations, an uncertainty factor of 10 is used by EPA and ATSDR. The value of 10 is derived from a threefold factor for differences in toxicokinetics and for threefold factor for toxicodynamics. To extrapolate from animals to humans and account for interspecies variability between humans and other mammals, an uncertainty factor of 10 is used by EPA and ATSDR, and as with intraspecies extrapolations, this 10-foldfactor is assumed to be associated with in toxicodynamics and toxicokinetics. An uncertainty

Table 24.2 Comparison of Less Serious Effects and Serious Effects

Less Serious




Reversible cellular changes


Necrosis, metaplasia, or atrophy



Clinically significant organ impairment

Delayed ossifciation

Visceral or skeletal abnormalities

Alteration in offspring weight

Cleft palate, fused ribs

Altered T-cellactivity

Necrosis inn immunologic components

Auditory disorders

Visual disorders

50% Reduction in offspring

Abnormal sperm





factor of 10 is used when a NOAEL derived from a subchronic study instead of a chronic study is used as the basis for a calculation of a chronic RfD (EPA only). Note that ATSDR does not perform this extrapolation but derive chronic and subchronic MRLs. An uncertainty factor of 10 is used in deriving an RfD or MRL from a LOAEL when a NOAEL is not available. It should be noted that there are no reference doses for dermal exposure, however when there is insufficient dermal absorption data, the EPA uses a default factor of 10% to estimate bioavailability for dermal absorption. A modifying factor ranging from 1 to 10 is included by EPA only to reflect a qualitative professional assessment of additional uncertainties in the critical study and in the entire data base for the chemical not explicitly addressed by preceding uncertainty factors.

Refinements of the RfC have utilized mechanistic data to modify the interspecies uncertainty factor of 10 (Jarabek, 1995). The reader should appreciate that with the inhalation route of exposure, dosimetric adjustments are necessary and can affect the extrapolations of toxicity data of inhaled agents for human health risk assessment. The EPA has included dosimetry modeling in RfC calculations, and the resulting dosimetric adjustment factor (DAF) used in determining the RfC is dependent on physiochemical properties of the inhaled toxicant as well as type of dosimetry model ranging from rudimentary to optimal model structures. In essence, the use of the DAF can reduce the default uncertainty factor for interspecies extrapolation from 10 to 3.16.

The 1996 Food Quality Protection Act (FQPA) now requires that an additional safety factor of 10 be used in the risk assessment of pesticides to ensure the safety of infants and children, unless the EPA can show that an adequate margin of safety is assured with out it (Scheuplein, 2000). The rational behind this additional safety factor is that infants and children have different dietary consumption patterns than adults and infants, and children are more susceptible to toxicants than adults. We do know from pharmacokinetics studies with various human pharmaceuticals that drug elimination is slower in infants up to 6 months of age than in adults, and therefore the potential exists for greater tissue concentrations and vulnerability for neonatal and postnatal effects. Based on these observations, the US EPA supports a default safety factor greater or less than 10, which may be used on the basis of reliable data. However, there are few scientific data from humans or animals that permit comparisons of sensitivities of children and adults, but there are some examples, such as lead, where children are the more sensitive population. It some cases qualitative differences in age-relatedsusceptibility are small beyond 6 months of age, and quantitative differences in toxicity between children and adults can sometimes be less than a factor of 2 or 3.

Much of the research efforts in risk assessment are therefore aimed at reducing the need to use these default uncertainty factors, although the risk assessor is limited by data quality of the chemical of interest. With sufficient data and the advent of sophisticated and validated physiologically based pharmacokinetic models and biologically based dose-responsemodels (Conolly and Butterworth, 1995), these default values can be replaced withscience-basedfactors. In some instances there may be sufficient data to be able to obtain distributions rather than point estimates.

24.3.2Derivation of Developmental Toxicant RfD

Developmental toxicity includes any detrimental effect produced by exposures during embryonic development, and the effect may be temporary or overt physical malformation. Adverse effects include death, structural abnormalities, altered growth, and


functional deficiencies. Maternal toxicity is also considered. The evidence is assessed and assigned a weight-of-evidencedesignation as follows: category A, category B, category C, and category D. The scheme takes into account the ratio of minimum maternotoxic dose to minimum teratogenic dose, the incidence of malformations and thus the shape of thedose-responsecurve or dose relatedness of the each malformation, and types of malformations at low doses. A range of uncertainty factors are also utilized according to designated category as follows: category A= 1–400,category B= 1–300,category C= 1–250,and category D= 1–100.Developmental RfDs are based a short duration of exposure and therefore cannot be applied to lifetime exposure.

24.3.3 Determination of RfD and RfC of Naphthalene with the NOAEL Approach

The inhalation RfC for naphthalene was 0.003 mg/m3, and this RfC was derived from a chronic(2-year)NTP inhalation study in mice using exposures of 0, 10, or 30 ppm (NTP, 1992). Groups of mice were exposed for 5 days a week and 6 hours a day. This study identified a LOAEL of 10 ppm. Adose-relatedincidence of chronic inflammation of the epithelium of the nasal passages and lungs was observed. This LOAEL concentration was normalized by adjusting for the6-hour-per-dayand5-day-per-weekexposure pattern. A LOAEL of 9.3 mg/m3 was obtained was derived by converting 10 ppm first to mg/m3 and thenduration-adjustedlevels for 6 h/day and 5 days/week for 103 weeks. An UF of 3000 was used, where 10 was for the interspecies (mice to humans) extrapolations, 10 for intraspecies variation in humans, 10 for using a LOAEL instead of a NOAEL, and 3 for database deficiencies.

The oral RfD for naphthalene was 0.02 mg/kg/day, and a study by Battelle (1980) was used to calculate the RfD. Decreased body weight was the most sensitive end point in groups of Fischer 344 rats given 0, 25, 50, 100, 200, or 400 mg/kg for 5 days/week for 13 weeks. These doses were also duration-adjustedto 0, 17.9, 35.7, 71.4, 142.9, and 285.7 mg/kg/day, respectively. The NOAEL fora > 10% decrease in body weight in this study was 71 mg/kg/day. The UF of 3000 was based on 10 for rats to humans extrapolation, 10 for human variation, 10 to extrapolate from subchronic to chronic, and 3 for database deficiencies including lack of chronic oral exposure studies.

24.3.4Benchmark Dose Approach

There are several problems associated with using the NOAEL approach to estimate RfDs and RfCs. The first obvious constraint is that the NOAEL must by definition be one of the experimental doses tested. Once this dose is identified, the rest of the dose-responsecurve is ignored. In some experimental designs where there is no identifiable NOAEL but LOAEL, thedose-responsecurve is again ignored, and the NOAEL is derived by application of uncertainty factors as described earlier. This NOAEL approach does not account for the variability in the estimate of the dose response, and furthermore experiments that test fewer animals result in larger NOAELs and thus larger RfDs and RfCs.

An alternative approach known as the benchmark dose (BMD) approach has been developed and implemented by risk assessors as an alternative to the NOAEL approach to estimate RfDs and RfCs. This approach is not constrained by experimental design